offcenter2005, post: 386109 wrote: If im unserstanding the question i just have my first behringer optical I/O ran to the first I/O on the main raydat card. Then the second behringer optical I/O ran to the second I/O on the other I/O on the main raydat card.

OK...so ADAT 1 and 2 on the RayDAT and your not using ADAT3/4 on the expansion card?
The reason I asked was because you have two separate ADAT converters that are converting analog audio into ADAT data independently and each one is transmitting it's own ADAT signal and clocking.
So channels 1-8 are one stream of data into the RayDAT and 9-16 the other.
So I was curious how the RayDAT sync's each stream independently but puts them onto the PCIe bus together.
It must be capable of reading and syncing each of the 4 ports and combining them together...which is pretty cool if that's how it's works...
In my mind this would be different if it was one 16 channel converter transmitting all 16 from the same box.
Maybe some ADAT expert here can elaborate!?

I was wondering the same thing. But since the raydat main card and the expansion card has two sets of I/Os on each i figured that was how i was supposed to set it up. I am extremely happy about the latency also. I guess the raydat has some of the lowest latency of any interface out. With it that low input monitoring with amp emulators and effects shouldn't have any noticeable lag. I know that when recording tracks you should always track dry but sometimes a little reverb and delay helps with tracking certain things. So far im very pleased. Any bugs that need worked out will come in timw when getting used to the interface. I would also like to hear from someone that is very familiar with this type of setup. Maybe im not running it correctly but its working so far.

I just got a reply on the rme site about which should be the master and slave. I guess that hooking up the first behringer as master and second as slave and also set the raydat to slave adat 1 is the best setup. I dont understand that at all because i thought that the raydat would be better as the master clock.

And was that answer from an employee of RME or just another forum user? I wouldn't trust a Behr*** clock for anything. I'd leave the RayDat as master and the other two units as slave or buy a Master clock and some BNC cables. You already ran a test this way without any issues.

TheJackAttack, post: 386163 wrote: And was that answer from an employee of RME or just another forum user? I wouldn't trust a Behr*** clock for anything. I'd leave the RayDat as master and the other two units as slave or buy a Master clock and some BNC cables. You already ran a test this way without any issues.

Im sure it wasnt an rme employee. But i figured id bring it up here to get some opinions. It is working great so far the way i have it. I was just wondering if i had it set the way i should for best performance. But since it is working im not messing with it until i get better converters. Thanks.

If you are looking for the "best" way to have the most stable clock, then that is with a master clock in either star array or in sequence with a termination cap (or termination switch on the last piece of gear). Provided your ADAT cables are high quality and don't get moved/bumped/bent much an ADAT based clock source is adequate for many things. I think two sets of ADAT ports is where I would stop with ADAT as clock source in the FWIW category. If I ran all four sets of ports then I'd get a master clock. You don't have to have the most expensive one on the market to get the benefit of stability either.

What you are saying is if i run all 4 I/Os on the raydat i should get an external WC? Arent clocks expensive? The big Ben is a lot i think right? But its considered one of the best though isnt it?

Sort of. Big Ben was one of the defacto standards many years ago. There are "higher" end than Ben currently offered like the Antelope Isochrone 10M or several offerings from Grimm. Then we would need to have an argument/discussion over how much benefit an external clock can ultimately make when it still interfaces with the in built clock circuit on the gear to which it's connected. (phew...lots of words for a dumb grunt) One of the problems with clocking from the ADAT is the deficiency of the cabling itself. Of all the types of clock connections, it is the most succeptible to damage and hence stability. For mobile rigs especially I would recommend BNC clocking rather than ADAT because of the constant plugging and unplugging, coiling and uncoiling. For permanent studio builds I still believe coax to be a better clock source but much of the cable stress is removed.

I have clocked from ADAT for my classical mobile rig on many occasions and had it work just fine. I still prefer coax as my clock source whether with external or RME's internal clock. Note: the clock on the Alesis HD24XR is NOT stable except at 48k for instance, so it is important to know your equipment. I trust RME's circuitry over Behr*** any day and twice on Sunday.

good then for what i am currently using it seems i have it set the way it should be. Ill worry about adding a external clock when i get more converters or upgrade my current ones to a better quality brand.

Basically. I would run it this way while I bought the coax to hookup the BNC wordclock. It would be better all around and it really doesn't cost that much I promise. Still use your RayDat as the master and then run to ADA8000 #1 with a T connector and then run on to ADA8000 #2. If you add 3 and 4 you just add a T onto the #2 unit and continue down the chain. The last in the line still has a T and then you put a termination cap on the unused connection to finish.

You can try the WC, but I still wouldn't be using the Berhingers as master...
I think you really need to run both ADA's like you have it setup right now and track the 16 channels simultaneously.
That is the only time you will have both fully packed data streams coming into the RayDAT and being recorded to Sonar (at least 5 minutes). That will be the proof that it's stable and everything stays in sync and if the ADAT glitches or looses sync or acts strange then that's when it will happen. Then I would look into doing the WC setup.

Just out of curiosity....are both the ADA's Sync Lock LED's lit on their front panels?
Does the RayDAT indicate ADAT is sync locked?

Both the behringer locked lights are on and in the hdsp settings the input status says adat1 sync 44.1 kHZ and adat2 sync 44.1 kHZ. Is there any other settings that i should mess with in sonar to optimize the raydat. I set it to ASIO and the at 44.1 kHZ and didnt mess with anything else.

No.

I set the buffer size to 32 samples and opened a 36 track song that i did for someone and it is loaded with plugins and it ran smooth without any pops or glitchs.

Great, but that isn't recording. Do test dj suggested.

Sent from my DROIDX using Tapatalk

No that is perfect!
The bullet proof test is the 16 track test....full blown max all ADAT channels going.
If it passes that stress test for at least 5 minutes of recording time and everything is accurately recorded with no dropouts and the playback is glitch free.....I would confidently say you can expect it will work flawlessly 24/7.
Make a note of the latency figures when you do that.
Even though you may never do 16....you are heavily dependent on the ADAT working properly as that is your only interface method and it will verify that everything stays together...

djmukilteo, post: 386187 wrote: No that is perfect!
The bullet proof test is the 16 track test....full blown max all ADAT channels going.
If it passes that stress test for at least 5 minutes of recording time and everything is accurately recorded with no dropouts and the playback is glitch free.....I would confidently say you can expect it will work flawlessly 24/7.
Make a note of the latency figures when you do that.
Even though you may never do 16....you are heavily dependent on the ADAT working properly as that is your only interface method and it will verify that everything stays together...

If i add more inputs would the stability at that sample rate be affected?

Also on that track my cpu meter is reading barley anything. Is that a good sign that it will stay stable with that many tracks and plugins?

offcenter2005, post: 386188 wrote: If i add more inputs would the stability at that sample rate be affected?

The sample rate (44.1khz) won't but the sample buffer setting will. That's why it's important to do that test.
The ability to stream 16 channel into your system and "print" those to your internal HDD (which is your maxed out setup) is all dependent on the buffer size which is a temporarily space storing the incoming bits before transferring them onto the HDD. It's how the incoming stream keeps up with the CPU, RAM and HDD. If it's too small and it fills up with data coming in and the HDD isn't ready to accept the buffer data it will crap out....dropouts, clicks, pops will start to occur. The bigger the buffer the less likely there will be a problem with that throughput. The bigger it is the more time it takes to fill before it transfers, the smaller the quicker...but if it's too quick somebody is not going to be ready and they have a problem.....they keep going but they will skip or drop some of the data. Somewhere in the middle is where you will find the perfect match of size and speed to keep up with each other. And as you know the latency increases with increasing the sample buffer size.
A 32 sample buffer in terms of size is small and fast and might work for one track or maybe even two or three, but 16 data packets coming in is a whole other story and may not work. Setting the buffer and finding that happy medium is all part of optimizing "your" system. The buffer can be changed depending on what your doing too...if your just playing back some tracks and recording one overdub and you want low latency you might find that 128 works best or 64 or 32, but when you set up 12 mics and 4 instruments and recording all that live at once...it will need to go up..
You'd have to find someone who has the "exact" same system you have in front of you (highly unlikely) to be able to figure out what works correctly!...so the only way for you to know is do it....there is no fixed setting that works on every machine...

djmukilteo, post: 386193 wrote: The sample rate (44.1khz) won't but the sample buffer setting will. That's why it's important to do that test.
The ability to stream 16 channel into your system and "print" those to your internal HDD (which is your maxed out setup) is all dependent on the buffer size which is a temporarily space storing the incoming bits before transferring them onto the HDD. It's how the incoming stream keeps up with the CPU, RAM and HDD. If it's too small and it fills up with data coming in and the HDD isn't ready to accept the buffer data it will crap out....dropouts, clicks, pops will start to occur. The bigger the buffer the less likely there will be a problem with that throughput. The bigger it is the more time it takes to fill before it transfers, the smaller the quicker...but if it's too quick somebody is not going to be ready and they have a problem.....they keep going but they will skip or drop some of the data. Somewhere in the middle is where you will find the perfect match of size and speed to keep up with each other. And as you know the latency increases with increasing the sample buffer size.
A 32 sample buffer in terms of size is small and fast and might work for one track or maybe even two or three, but 16 data packets coming in is a whole other story and may not work. Setting the buffer and finding that happy medium is all part of optimizing "your" system. The buffer can be changed depending on what your doing too...if your just playing back some tracks and recording one overdub and you want low latency you might find that 128 works best or 64 or 32, but when you set up 12 mics and 4 instruments and recording all that live at once...it will need to go up..
You'd have to find someone who has the "exact" same system you have in front of you (highly unlikely) to be able to figure out what works correctly!...so the only way for you to know is do it....there is no fixed setting that works on every machine...

What are the bennifits of lower sample rate? I know it lowers the latency but is that really that important unless you are monitoring the effects as you record?

Sample rate or sample buffer setting?
Sample Rate is the 44.1kHz or 48kHz number. That is the speed at which sample conversion of your analog signal being fed into your ADA's is being converted into ADAT data packets. Your Sonar project needs to match that number and your done. 44.1khz is the default standard, 48kHz is just a little faster, and you can argue all you like about one being better than the other but it has nothing to do with "faster".

The sample "buffer" number (32, 64, 128 etc) is the buffer size space setup by the RayDAT so there is temporary storage space for the packets of data that are coming from your ADA's into the RayDAT. The smaller the buffer "size" the quicker the buffer will fill with data and then try to dump that buffer data to your HDD. Depending on the overall throughput of your computer system the buffer will either lag or lead the CPU, RAM and HDD depending on it's size to complete that task. This of course is all happening in microseconds or nanoseconds. So there's a certain point where they will work smoothly together and a certain point where they will not depending on how much data is being processed. You adjust the buffer size by increasing or deceasing it to find that "sweet" spot.

djmukilteo, post: 386196 wrote: Sample rate or sample buffer setting?
Sample Rate is the 44.1kHz or 48kHz number. That is the speed at which sample conversion of your analog signal being fed into your ADA's is being converted into ADAT data packets. Your Sonar project needs to match that number and your done. 44.1khz is the default standard, 48kHz is just a little faster, and you can argue all you like about one being better than the other but it has nothing to do with "faster".

The sample "buffer" number (32, 64, 128 etc) is the buffer size space setup by the RayDAT so there is temporary storage space for the packets of data that are coming from your ADA's into the RayDAT. The smaller the buffer "size" the quicker the buffer will fill with data and then try to dump that buffer data to your HDD. Depending on the overall throughput of your computer system the buffer will either lag or lead the CPU, RAM and HDD depending on it's size to complete that task. This of course is all happening in microseconds or nanoseconds. So there's a certain point where they will work smoothly together and a certain point where they will not depending on how much data is being processed. You adjust the buffer size by increasing or deceasing it to find that "sweet" spot.

I meant to say sample buffer size. As i said before i have it set to 64 and it seems to do fine when recording the 16 channels and let it run for around 6 minutes. Is there anything i can do after setting that to help stabalize the system to run with the lower buffers settings? I dont really see myself needing to ever use more than 36 inputs at once, even 16 in most cases.

If you have 16 analog channels recording simultaneously with a 64 buffer, you have a stellar system!
What was the latency figure?
I don't see where you will ever have more channels than that with the 2 ADA's you have now.
If the latency is acceptable to your ears then leave it.
If you want to experiment and see where the exact limits are you can go down or up a size, run the same test again and see if it starts crappin out...you'll see or hear it, maybe a good test to recognize what it sounds like.
It's just a balance, if it doesn't crap out, once your there that's it...
If you were to add 2 more ADAT converters and have 32 channels simultaneously (which would be the max ADAT for the RayDAT to handle) then you'll have to run a test with all that running and see what the results are and adjust accordingly..

I probably wont need to add another ADA, I think 16 inputs will do most recording jobs. If anything i would add 1 more if needed. But before that I would rather upgrade quality than quantity. At a buffer size of 64 im getting 3.6 milliseconds roundtrip the way its setup. I can hear it but its not that bad. When i run the DPC latency test how do i do that?

I don't think you could do any better than that....I'm surprised you can hear 4msec!...most people can only hear 15msec or above....
The DPC tool just has a start button there and it will start scrolling across the little screen in the app until you tell it to reset. It will report if your system is good and that is typically when the bars are all green.
Other instructions and uses should be on that website.

I dont hear any latency when plugged straight in, but when using the input monitoring switch i hear a very slight delay. But it is really small.